▸ The Stop Killing Games initiative crossed 1.4M verified EU signatures, forcing the Commission to formally respond to a law banning publishers from rem...
▸ Tailscale published a rare vendor post-mortem admitting its network didn't stop the Hugging Face intrusion — because that was never the job.
▸ DoD counterintelligence quietly elevated Israel to the same threat tier as China, Russia, Iran, and North Korea — a category reserved for adversaries ...
▸ Meta's chatbot didn't have a vulnerability — the password-reset endpoint it called did, and the AI just gave attackers a friendly UI to hit it.
▸ A fourth suspicious repo — Parcle-AI/parcle-memory at 243 stars — has joined chokepoint-atlas, reg-factory, and ConiferKit/sage on GitHub Trending in ...
▸ Linear feels instant because almost nothing in the UI waits on a network round-trip — the client owns the data and the server is a sync target, not a ...
▸ Interpol's 2026 African Cyberthreat Assessment attributes more than 50% of cybercrime on the continent to AI-assisted tooling.
▸ An Illinois cop used Flock's ALPR network to track a driver across state lines, then invented a pretext to search his car for weed.
▸ The BBC piece formalizes what feed data has shown for two years: the median post you see is from a stranger optimizing for a trend, not a friend.
▸ Qwen3.7-Max dropped open weights while OpenAI, Anthropic, and Google ship behind metered APIs and $200/mo tiers.
▸ The Hamburg ruling reframed Google's AI Overviews as first-party speech, killing the 'we just summarized the web' defense for any generative product s...
▸ A developer is documenting a from-scratch software rasterizer with no GPU, no shaders, no z-buffer — the actual 1993 toolkit.
▸ An AI agent ran unattended through Fedora's infrastructure, generating low-quality patches, bug reports, and reviews until maintainers shut it down.
▸ Waymo dropped the waitlist in Dallas — anyone with the app can now hail a fully driverless ride.
▸ Researchers at the American Diabetes Association's 2026 Scientific Sessions were ejected after distributing PDF reprints of peer-reviewed papers to at...
▸ Mohkohn ripped out the SPA, shipped server-rendered HTML, and watched daily users double inside a week.
▸ OpenAI published a rare public blog post accusing Apple of anti-competitive App Store behavior that hurts AI apps and users.
▸ Sean Goedecke's viral post argues LLMs are a multiplier on existing skill, not a replacement for it — and the HN thread agrees.
▸ A 'critical' SQLite CVE turned out to be an LLM hallucination that made it through the official CVE process before JFrog's researchers debunked it.
▸ A senior engineer's bearblog post arguing LLMs are degrading both the job and the craft hit 320 on HN, triggering the largest 'where is this going' th...